1566
Cyclopropane Formation via Radical-Induced Methylene Transfer to Olefins'
si.,.. We wish to report a new type of free-radical reaction, transfer of a methylene group to an olefin to give a cyclopropane.
In view of the present lack of mechanistic information, we do not feel justified in invoking reaction paths more involved than reaction between iodomethyl radical and olefin to give a cyclopropane. This process must proceed with loss of iodine either before,%during,4 or after6 carbon-carbon bond formation. Our observations are clearly relevant to the question of the mechanismlo of the photolytic reaction of methylene iodide with cyclohexene,8,11 3-l1exene,~ and 2butene8 to give the corresponding cyclopropanes.
0.13 0.078
Q.047
0.024
0.004 0.033
(2) Strong evidence against this route is provided by a crudely estimated endothermicity of 70-85 kcal/mole for the reaction . C H d :CHz I . . Our formulation of this possibility must remain vague because we are unable to predict with confidence the relative amounts of singlet and triplet :CHz that would be initially produced by loss of iodine from . C H L It should be noted that the reactions .CCh C1. :CCh3" and .CF3 + :CFZ F.'b have been invoked. (3) (a) G . P. Semeluk and R. B. Bernstein, J. A m . Chem. SOC., 79, 46 (1957); G. Archer and C. Hinshelwood, Proc. Roy. SOC.(London), A261, 293 (1961); M. Seakins, ibid., A274, 413 (1963). However, see also: A. E. Shilov and R. D. Sabirova, Zh. Fiz. Khim., 33, 1365 (1959); 34, 860 (1960); J. W. Engelsma, Rec. Trac. Chim., 84, 187 (1965); D. J. Clark and J. M. Tedder, Trans. Faraday SOC.,62, 393, 399, 405 (1966). (b) J. W. Hodgins and R. L. Haines, Can. J . Chem., 30, 473 (1952). However, see E. Tschuikow-Roux, J . Chem. Phys., 43, 2251 (1965). (4) This is an unprecedented (in free-radical chemistry), but not a priori unreasonable, process. An analog in carbonium ion chemistry would be the proposed T route to "nonclassical" cations, an idea which does not lack supporters. Similar processes have been proposed as models for the reaction of carbenoids with olefins to give cyclopropanes,b the Simmons-Smith reaction,' and the photolytic reaction of methylene iodide with olefins.8 We have obtained preliminary results under conditions such that there is no cis-trans interconversion among the olefins or the cyclopropanes, which indicates that the reaction of cis- and trans4-octenes to give 1,2-dipropylcyclopropane is not stereospecific, both cis- and trans-4-octenes leading to a trans : c i s ratio of 1,2-dipropylcyclopropane of -20. This observation cannot be readily accommodated to a mechanism whereby ring formation and iodine loss are concerted. (5) This possibility is very reasonable in view of our recent radicalinduced conversion of 1,3-diiodopropane to cyclopropane. 8 In this communication we also mentioned the above conversion of 1,l-diphenylethylene to 1,l-diphenylcyclopropane. (6) G. L. Closs and R. A. Moss, J . A m . Chem. SOC.,86,4042 (1964); G . L. Closs and L. E. Closs, Angew. Chem. Intern. Ed. Engl., 1, 334 (1962). (7) H. E. Simmons and R. D. Smith, J . Am. Chem. SOC.,80, 5323 (19 58). (8) D. C. Blomstrom, K. Herbig, and H. E. Simmons, J . Org. Chem., 30, 959 (1965). (9) L. Kaplan, J . Am. Chem. SOC.,89, 1753 (1967). (10) It was concluded* that the intermediacy of iodomethyl radicals is unlikely. (1 1) R. C. Neuman, Jr., and R. S. Wolcott, Tetrahedron Letters, 6267 (1966). Leonard Kaplan Department of Chemistry, Unioersity of Chicago Chicago, Illinois 60637 Receiued May 5, 1967
0.062 Trace 0.34
Influence of Metallic Cation in Carbonyl Addition Reactions of Certain Nitriles and Sulfones with Benzophenone
Results of a preliminary study involving a few olefins are given in Table I. Table I1 contains more complete data for the reaction of 1,l-diphenylethylene.
+
-+
+
Table I. Peroxide-Induced Reaction of Methylene Iodide with Olefins
% yielda of the corresponding cyclopropane -(t-BuO)n4PhCOz)z166" 135" 115" 81" 1-Octene P hCH=CH* Ph?C=CH2 Cyclohexene cis-Stilbenec traits-Stilbenec cis-4-Octened trans-4-Octened
38 17 14
17 5 b
absence of methylene iodide and not all of which have been iilexified. Material balances on di-t-butyl peroxide derived products and iodine are good, however.
(1) This work was supported by a Frederick Gardner Cottrell grant from the Research Corporation.
Journal of the American Chemical Society / 89:17
Sir : It has previously been shown that sodiophenylacetonitrile (1', M = Na), prepared from phenylacetonitrile and sodium amide in liquid ammonia, can be alkylated' and acylated2 in this medium to form I1 and 111, respectively. M
R
CoHjCHCN I
C&lb!XCN
I'
I1
RC===O CsHsLHCN
111
(1) A. C.Cope, H . L. Holmes, and H. 0. House, Org. Reactions, 9 , 107 (1957); W. G. Kenyon, E. M. Kaiser, and C. R. Hauser, J . Org. Chem., 30, 4135 (1965). (2) R. Levine and C. R. Hauser, J . Am. Chem. SOC.,68, 760 (1946).
August 16,1967
4567 Sodionitrile I’ has also been condensed with benzophenone in refluxing ether to form the a,P-unsaturated nitrile IV, which presumably arose through intermediate sodio adduct V’ (M = Na); the neutral adduct V was not isolated. In fact, hydroxy nitrile V has not been reported previously.
Apparently, the equilibrium of the addition reaction of sodio or lithio nitrile I’ with benzophenone in liquid ammonia is on the side of I’ and the ketone. Thus, we have observed that, not only are phenylacetonitrile and benzophenone recovered on treating the sodio or lithio nitrile I’ with the ketone and neutralizing the reaction mixture inversely, but the nitrile and ketone are also obtained on treating the hydroxy nitrile V (prepared as described below) with sodium amide under similar conditions. We now wish to report that the hydroxy nitrile V can be prepared from phenylacetonitrile and benzophenone if M of the intermediate I’ is a metal possessing a sufficiently greater coordinating (chelating) capacity than sodium and the reaction is effected in an inert solvent. Thus, adduct V was obtained in yields of 406 3 z when M of intermediate I’ was MgBr4 or an aluminum chloride moiety, and in 11 yield when M was lithium. One procedure is illustrated in Scheme I.
z
Scheme I
(CHa)zCHMgBr
CsHsCHzCN
I’ (M = MgBr)
ether
I
i(CoHd*CO
inverse
V
t-------V’(M neutralization
= MgBr)
Other procedures involved preparation of lithiophenylacetonitrile (1’, M = Li) from phenylacetonitrile and n-butyllithium in tetrahydrofuran-hexane and treatment of it with either magnesium bromide followed by benzophenone or preferably with a preformed 2 : l complex of magnesium bromide-benzophenone in THF. The aluminum chloride type intermediate I’ was prepared by a procedure analogous to the latter one. Similarly, the magnesium bromide salt VI’ of benzyl phenyl sulfone and the dimagnesium bromide salt VII” of dibenzyl sulfone (M = MgBr), prepared from the sulfone and isopropylmagnesium bromide or from the sulfone, n-butyllithium, and magnesium bromide, underwent carbonyl addition reactions with benzophenone to form hydroxy sulfones VI11 and IX in yields of 50-52 and 42-56%, respectively. In contrast to these results, the lithium, sodium, or potassium salts VI’ and VII” (M = Li, Na, or K) failed to afford these products on treatment with the ketone. (3) S. Wawzonek and E. M. Smolin, “Organic Syntheses,” Coll. Vol. IV, John Wiley and Sons, Inc., New York, N. Y., 1963, p 387. (4) This reagent (I’, M = MgBr) has a formal similarity to the Ivanov reagent, CaHsCH(MgCI)COONa, which is known to react with carbonyl compounds; however, disodio- or dilithiophenyl acetate are preferable since they give better yields of 8-hydroxy acids. See P. J. Hamrick, Jr., and C. R. Hauser, J . A m . Chem. SOC.,82, 1957 (1960).
M
M
M
The structures of the hydroxy nitrile V and the hydroxy sulfones VI11 and IX were supported by analyses, infrared and nmr spectra, and dehydration to afford the corresponding a,&unsaturated nitrile IV and sulfones X and XI, respectively. (CsH dzc
(C6Hs)zC
ll
CsH~cS02CsHs
X
/I
C~HSCSOZCHZC~H~
XI
However, the corresponding benzophenone adducts of malonic ester and acetylacetone failed to be isolated on preparing their respective magnesium bromide derivatives and treating them with the ketone under similar conditions. Instead, only starting materials were recovered. These results support the hypothesis that, although the main driving force in such carbonyl addition reactions of carbanions is considered to be the formation of a weaker base or nucleophile5 (Le., the alkoxide type ion, as in V’), an additional driving force can arise through coordination (chelation) within the product.6-8 This additional factor appears to be ineffective, however, when the difference between the strengths of the starting carbanion and the resulting alkoxide type ion is relatively great, since even the magnesium salts of the more weakly basic or nucleophilic carbanions failed to afford adducts (see above). Studies are in progress to ascertain the relative importance of weaker base formation, chelation, and steric factorss and to determine the generality and usefulness of the method in synthesis. Also, the mechanism of such addition reactions is being investigated. Acknowledgment. Acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for partial support of this research at the University of Missouri, and to the National Science Foundation for partial support at Duke University. ( 5 ) This hypothesis is supported by the fact that cleavage of such a benzophenone adduct occurs in the presence of a catalytic amount of base to form the thermodynamically more stable neutral starting materials; see P. J. Hamrick and C. R. Hauser, J . Am. Chem. SOC., 81, 3146 (1959). (6) See C. R. Hauser and W. R. Dunnavant, J . Org. Chem., 25, 1300 (1 960). (7) A similar shift in equilibrium has been observed in the reactions of nitroparaffins and carbon dioxide in the presence of magnesium and aluminum ions; see H. L. Finkbeiner and M. Stiles, J . A m . Chem. SOC., 85, 616 (1963). (8) A steric factor may also be involved since certain adducts appear to be obtained more readily from benzaldehyde than from benzophenone. Edwin M. Kaiser
Department of Chemistry, University of Missouri Columbia, Missouri 65201 Charles
R. Hauser
Department of Chemistry, Duke University Durham, North Carolina 27706 Receiced June 27, 1967 Communications to the Editor
